Family background, education and employment in urban Ethiopia

A visitor to Addis Ababa is likely to remark not merely on the variety of urban employment but also on the fact that some employments such as shoe-shining seem so slight as to hardly be employment at all. This is a commonplace for any developing country and most economists have concluded that the labour market is segmented into formal and informal sectors; earnings in different sectors differ for workers of equal productivity and that entry into formal employment is rationed. Some have gone further and suggested that the informal sector is not homogenous and that it could be divided into an intermediate sector where entry is also restricted due to limited access to capital and skills and a residual free entry sector that includes shoe-shine boys'. Associated with such diversity, is the idea of differential access to jobs and that often access depends on kinship and other networks'. More casual observation by the same visitor might lead him to conclude that in a country with a very low enrolment in primary or secondary education and formal employment confined to the public sector, stringent selection must be at work restraining both education and employment. There is no excess supply of schooling; rationing is likely both through location of schools as well as through access by family background or through selective examinations. Similarly, access to employment could be limited by family connections or by selective recruitment on the part of firms. One possibility is that education is used as a screening device by firms particularly in the formal sector so that more highly educated workers are always preferred. This must result in a mis-match between occupations and skills and in the mis-allocation of labour within the formal sector. It also has an impact on the composition of the unemployed; if, as is likely, the educated are also from better-off families it will skew the distribution of the unemployed towards the better-off, educated who can afford to wait for formal sector employment. The stand taken in this paper is that segmentation is too slippery a concept to be useful or warrant a formal test but that the idea that labour cannot be aggregated across sectors is sensible in a developing country context'. The point of departure of the paper is the treatment of unemployment. I argue that it ought to be modelled together with the allocation of labour into the formal and informal sectors. The focus is on the process of labour allocation; to discover to what extent alternative earnings prospects as distinct from background or financial constraints influence the choice of occupation. Outcomes that result from such allocation may be unsatisfactory because access to the better jobs is unequal. Some kinds of households may thus be more likely to face poverty-level incomes than others. In particular, the importance of family background in determining both participation in employment and incomes is stressed. The impact of family background on earnings is usually seen as capturing the quality of the learning environment or proxying omitted ability. It has been suggested that it could also be interpreted as social stratification. For instance, other things being equal, more educated parents might be able to assure access to modern sector employment. Thus, in an imperfect labour market the direct effect of parents' education on the earnings of their children would not necessarily correspond to higher productivity of the children. The key issue therefore is whether it is possible to test the hypothesis that the direct effect of parents' education on childrens' education truly reflects an increase in productivity or not and the last part of this paper is devoted to this issue. The paper considers the urban labour market in Ethiopia in 1990. This is a year of peculiar interest for it is the last year in power of the previous socialist government which had been in power for 18 years. However, it is also a typical year in that the supply of educational facilities in urban areas stayed the same over the previous decades. In short, the educational structure did not change rapidly (or indeed at all) in this period and so warrants the cross-section basis of the analysis in this paper. Neither did labour market conditions change dramatically. The following section presents a short-run microeconomic model of labour supply and earnings. The purpose is to build a description of the private sector labour market as well as the unemployed. The results of the estimation are used to build a complete picture of the role of family background in determining educational attainment and earnings in the public and private sectors of the Ethiopian labour market. In passing, the paper also examines the robustness of the usual methods employed in estimating models of polychotomous choice and selection.